Mixed messages on new bleeding data with dabigatran

March 15, 2013

San Francisco, CA - A new analysis of the US Food and Drug Administration (FDA) Mini-Sentinel database has given reassuring results on the rates of gastrointestinal (GI) and intracranial bleeds with dabigatran (Pradaxa, Boehringer Ingelheim) [1].

However, a separate analysis of dabigatran bleeding reports submitted to the FDA, presented at the recent American College of Cardiology (ACC) 2013 Scientific Sessions, has suggested a much higher case fatality rate than that reported in the major clinical trials of the drug [2].

Mini-Sentinel analysis

The Mini-Sentinel analysis, published online in the New England Journal of Medicine on March 13, 2013, was conducted by three FDA employees in response to the high numbers of bleeding reports with dabigatran when it was first approved in the US.

The researchers, with first author Dr Mary Ross Southworth (FDA Center for Drug Evaluation and Research), examined the Mini-Sentinel database, which monitors routinely collected electronic healthcare data, for the period from October 19, 2010 (the date of dabigatran approval), to December 31, 2011, to identify inpatient diagnosis codes for intracranial and GI hemorrhages associated with new use of dabigatran or warfarin.

Results showed that bleeding rates associated with dabigatran use during that period did not appear to be higher than those associated with warfarin.

Intracranial and GI bleeds in new users of dabigatran and warfarin for AF

End point
n (events)
Events per 100 000 days at risk
n (events)
Events per 100 000 days at risk
GI h emorrhage
10 599 (16)
43 541 (160)
Intracranial h emorrhage
10 587 (8)
43 594 (109)

Noting that there are limitations to the Mini-Sentinel analysis, including lack of adjustment for confounding variables and lack of a detailed medical record review, the authors nevertheless say they believe these data are reassuring.

They conclude that the large number of reported cases of bleeding associated with dabigatran was probably an example of stimulated reporting due to publicity surrounding the launch of the drug and a tendency to preferentially report adverse events with new drugs.

They add that further analysis of the Mini-Sentinel and other claims databases is ongoing, as is routine postmarketing surveillance.

Adverse events reported to FDA

The other analysis, presented at the ACC meeting by Dr Kevin W McConeghy (University of Illinois at Chicago), examined publicly available data from adverse-event reports submitted to the FDA on dabigatran.

McConeghy stated that the aims of his study were to describe the population of dabigatran users experiencing adverse events; estimate the risk of death among FDA bleeding reports associated with dabigatran; and determine a lower-bound estimate of dabigatran-related bleed fatalities in the US.

For the study, the researchers examined reports of bleeding with dabigatran or warfarin submitted to the FDA between January 1, 2010, and June 30, 2012.

Adverse drug reactions relating to warfarin or dabigatran reported to FDA

End point
Total adverse events report ed (n)
12 581
Bleeding e vents , n (%)
590 (13)
2453 (19)
% f emale
Age ( y )
Weight (kg)

McConeghy pointed out that the mean age of patients who had an adverse event with dabigatran reported to the FDA (75) was higher than those in the RE-LY trial (71), and more patients who had FDA reported events were female (48%) than included in RELY (36.7%).

"This suggests the drug is being used in a different profile of patients from that in the RE-LY trial," he said.

Of the 2453 dabigatran bleeding adverse events reported to the FDA, 393 (16%) were fatal, almost double the case fatality rate of patients who bled in the five phase 3 trials of the drug, he noted.

Bleeding events reported to the FDA

Bleeding event
Total bleeds (n)
GI bleeds, n (%)
162 (27%)
1352 (55%)
Intracranial bleed , n (%)
69 (12%)
280 (11%)
Bleed fatalities , n (%)
47 (8%)
393 (16%)

Thirty-day mortality after the first major bleed in five phase 3 trials was 13% for warfarin-treated patients and 9% for dabigatran-treated patients.

Using the FDA adverse-event-report data and a previously published estimate that there were 232 000 patient-years of exposure to dabigatran in 2011 in the US, they calculated a fatal event rate of 95.3 per 100 000 person-years. McConeghy said this could be considered a lower-bound estimate, because not all adverse events are reported. He worked out that if fewer than one in three adverse bleeding events were reported to the FDA, this fatality rate may signal an increased risk of bleed fatalities compared with that seen in RE-LY.

But he cautioned that the figures used in this analysis suffered from reporting bias, accuracy, and duplication problems and so should not be used to draw any definite conclusions.


Comments on Medscape are moderated and should be professional in tone and on topic. You must declare any conflicts of interest related to your comments and responses. Please see our Commenting Guide for further information. We reserve the right to remove posts at our sole discretion.